Skip to contentSkip to conversation

powerpc/book3s/mm: Update Oops message to print the correct translation in use

PATCH Applied Aneesh Kumar K.V 5 participants 8 messages Jul 11, 2019 – Nov 14, 2019 PowerPC

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 powerpc/book3s/mm: Update Oops message to print the correct translation in use

Avoids confusion when printing Oops message like below

Faulting instruction address: 0xc00000000008bdb4

 Oops: Kernel access of bad area, sig: 11 [#1]
 LE PAGE_SIZE=64K MMU=Radix MMU=Hash SMP NR_CPUS=2048 NUMA PowerNV

This was because we never clear the MMU_FTR_HPTE_TABLE feature flag even if we run with radix translation. It was discussed that we should look at this feature flag as an indication of the capability to run hash translation and we should not clear the flag even if we run in radix translation. All the code paths check for radix_enabled() check and if found true consider we are running with radix translation. Follow the same sequence for finding the MMU translation string to be used in Oops message.

---

Changes from V1:
* Don't clear the HPTE feature flag.

 arch/powerpc/kernel/traps.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)
arch/powerpc/kernel/traps.c +11 -4
index 11caa0291254..b181d6860f28 100644
@@ -250,15 +250,22 @@ static void oops_end(unsigned long flags, struct pt_regs *regs,
 }
 NOKPROBE_SYMBOL(oops_end);
 
+static char *get_mmu_str(void)
+{
+	if (early_radix_enabled())
+		return " MMU=Radix";
+	if (early_mmu_has_feature(MMU_FTR_HPTE_TABLE))
+		return " MMU=Hash";
+	return "";
+}
+
 static int __die(const char *str, struct pt_regs *regs, long err)
 {
 	printk("Oops: %s, sig: %ld [#%d]\n", str, err, ++die_counter);
 
-	printk("%s PAGE_SIZE=%luK%s%s%s%s%s%s%s %s\n",
+	printk("%s PAGE_SIZE=%luK%s%s%s%s%s%s %s\n",
 	       IS_ENABLED(CONFIG_CPU_LITTLE_ENDIAN) ? "LE" : "BE",
-	       PAGE_SIZE / 1024,
-	       early_radix_enabled() ? " MMU=Radix" : "",
-	       early_mmu_has_feature(MMU_FTR_HPTE_TABLE) ? " MMU=Hash" : "",
+	       PAGE_SIZE / 1024, get_mmu_str(),
 	       IS_ENABLED(CONFIG_PREEMPT) ? " PREEMPT" : "",
 	       IS_ENABLED(CONFIG_SMP) ? " SMP" : "",
 	       IS_ENABLED(CONFIG_SMP) ? (" NR_CPUS=" __stringify(NR_CPUS)) : "",
Signature · 1 line
2.21.0
reviewer Acked-by on v2 1/1
#1 lore raw

Aneesh Kumar K.V's on July 12, 2019 12:58 am:

Quoted 15 lines

Avoids confusion when printing Oops message like below

Faulting instruction address: 0xc00000000008bdb4

 Oops: Kernel access of bad area, sig: 11 [#1]
 LE PAGE_SIZE=64K MMU=Radix MMU=Hash SMP NR_CPUS=2048 NUMA PowerNV

This was because we never clear the MMU_FTR_HPTE_TABLE feature flag even if we run with radix translation. It was discussed that we should look at this feature flag as an indication of the capability to run hash translation and we should not clear the flag even if we run in radix translation. All the code paths check for radix_enabled() check and if found true consider we are running with radix translation. Follow the same sequence for finding the MMU translation string to be used in Oops message.

Quoted 5 lines

---

Changes from V1:
* Don't clear the HPTE feature flag.

Thanks,

Quoted 39 lines
 arch/powerpc/kernel/traps.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)
arch/powerpc/kernel/traps.c +11 -4
index 11caa0291254..b181d6860f28 100644
@@ -250,15 +250,22 @@ static void oops_end(unsigned long flags, struct pt_regs *regs,
 }
 NOKPROBE_SYMBOL(oops_end);
 
+static char *get_mmu_str(void)
+{
+	if (early_radix_enabled())
+		return " MMU=Radix";
+	if (early_mmu_has_feature(MMU_FTR_HPTE_TABLE))
+		return " MMU=Hash";
+	return "";
+}
+
 static int __die(const char *str, struct pt_regs *regs, long err)
 {
 	printk("Oops: %s, sig: %ld [#%d]\n", str, err, ++die_counter);
 
-	printk("%s PAGE_SIZE=%luK%s%s%s%s%s%s%s %s\n",
+	printk("%s PAGE_SIZE=%luK%s%s%s%s%s%s %s\n",
 	       IS_ENABLED(CONFIG_CPU_LITTLE_ENDIAN) ? "LE" : "BE",
-	       PAGE_SIZE / 1024,
-	       early_radix_enabled() ? " MMU=Radix" : "",
-	       early_mmu_has_feature(MMU_FTR_HPTE_TABLE) ? " MMU=Hash" : "",
+	       PAGE_SIZE / 1024, get_mmu_str(),
 	       IS_ENABLED(CONFIG_PREEMPT) ? " PREEMPT" : "",
 	       IS_ENABLED(CONFIG_SMP) ? " SMP" : "",
 	       IS_ENABLED(CONFIG_SMP) ? (" NR_CPUS=" __stringify(NR_CPUS)) : "",
Signature · 1 line
2.21.0
Reviewed-by on v2 1/1
#2 lore raw
In reply to Aneesh Kumar K.V

Le 11/07/2019 à 16:58, Aneesh Kumar K.V a écrit :

Quoted 15 lines

Avoids confusion when printing Oops message like below

Faulting instruction address: 0xc00000000008bdb4

  Oops: Kernel access of bad area, sig: 11 [#1]
  LE PAGE_SIZE=64K MMU=Radix MMU=Hash SMP NR_CPUS=2048 NUMA PowerNV

This was because we never clear the MMU_FTR_HPTE_TABLE feature flag even if we run with radix translation. It was discussed that we should look at this feature flag as an indication of the capability to run hash translation and we should not clear the flag even if we run in radix translation. All the code paths check for radix_enabled() check and if found true consider we are running with radix translation. Follow the same sequence for finding the MMU translation string to be used in Oops message.

Quoted 1 line

Looks good.

Quoted 40 lines

---

Changes from V1:
* Don't clear the HPTE feature flag.

  arch/powerpc/kernel/traps.c | 15 +++++++++++----
  1 file changed, 11 insertions(+), 4 deletions(-)
arch/powerpc/kernel/traps.c +11 -4
index 11caa0291254..b181d6860f28 100644
@@ -250,15 +250,22 @@ static void oops_end(unsigned long flags, struct pt_regs *regs,
  }
  NOKPROBE_SYMBOL(oops_end);
 
+static char *get_mmu_str(void)
+{
+	if (early_radix_enabled())
+		return " MMU=Radix";
+	if (early_mmu_has_feature(MMU_FTR_HPTE_TABLE))
+		return " MMU=Hash";
+	return "";
+}
+
  static int __die(const char *str, struct pt_regs *regs, long err)
  {
  	printk("Oops: %s, sig: %ld [#%d]\n", str, err, ++die_counter);
 
-	printk("%s PAGE_SIZE=%luK%s%s%s%s%s%s%s %s\n",
+	printk("%s PAGE_SIZE=%luK%s%s%s%s%s%s %s\n",
  	       IS_ENABLED(CONFIG_CPU_LITTLE_ENDIAN) ? "LE" : "BE",
-	       PAGE_SIZE / 1024,
-	       early_radix_enabled() ? " MMU=Radix" : "",
-	       early_mmu_has_feature(MMU_FTR_HPTE_TABLE) ? " MMU=Hash" : "",
+	       PAGE_SIZE / 1024, get_mmu_str(),
  	       IS_ENABLED(CONFIG_PREEMPT) ? " PREEMPT" : "",
  	       IS_ENABLED(CONFIG_SMP) ? " SMP" : "",
  	       IS_ENABLED(CONFIG_SMP) ? (" NR_CPUS=" __stringify(NR_CPUS)) : "",
#3 lore raw
In reply to Aneesh Kumar K.V
Quoted 30 lines · Aneesh Kumar K.V

Avoids confusion when printing Oops message like below

Faulting instruction address: 0xc00000000008bdb4

 Oops: Kernel access of bad area, sig: 11 [#1]
 LE PAGE_SIZE=64K MMU=Radix MMU=Hash SMP NR_CPUS=2048 NUMA PowerNV

This was because we never clear the MMU_FTR_HPTE_TABLE feature flag even if we run with radix translation. It was discussed that we should look at this feature flag as an indication of the capability to run hash translation and we should not clear the flag even if we run in radix translation. All the code paths check for radix_enabled() check and if found true consider we are running with radix translation. Follow the same sequence for finding the MMU translation string to be used in Oops message.

---

Changes from V1:
* Don't clear the HPTE feature flag.

 arch/powerpc/kernel/traps.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)
arch/powerpc/kernel/traps.c +0 -0
index 11caa0291254..b181d6860f28 100644
@@ -250,15 +250,22 @@ static void oops_end(unsigned long flags, struct pt_regs *regs,
 }

arch/powerpc/kernel/traps.c static void oops_end(unsigned long flags, struct pt_regs *regs, · line 260

Quoted 10 lines
arch/powerpc/kernel/traps.c +8 -0
 NOKPROBE_SYMBOL(oops_end);
 
+static char *get_mmu_str(void)
+{
+	if (early_radix_enabled())
+		return " MMU=Radix";
+	if (early_mmu_has_feature(MMU_FTR_HPTE_TABLE))
+		return " MMU=Hash";
+	return "";
+}

We don't change MMU once we're up, so just do this logic once and stash it into a static string, rather than rechecking on every oops.

cheers

Quoted 16 lines
arch/powerpc/kernel/traps.c +2 -4
 static int __die(const char *str, struct pt_regs *regs, long err)
 {
 	printk("Oops: %s, sig: %ld [#%d]\n", str, err, ++die_counter);
 
-	printk("%s PAGE_SIZE=%luK%s%s%s%s%s%s%s %s\n",
+	printk("%s PAGE_SIZE=%luK%s%s%s%s%s%s %s\n",
 	       IS_ENABLED(CONFIG_CPU_LITTLE_ENDIAN) ? "LE" : "BE",
-	       PAGE_SIZE / 1024,
-	       early_radix_enabled() ? " MMU=Radix" : "",
-	       early_mmu_has_feature(MMU_FTR_HPTE_TABLE) ? " MMU=Hash" : "",
+	       PAGE_SIZE / 1024, get_mmu_str(),
 	       IS_ENABLED(CONFIG_PREEMPT) ? " PREEMPT" : "",
 	       IS_ENABLED(CONFIG_SMP) ? " SMP" : "",
 	       IS_ENABLED(CONFIG_SMP) ? (" NR_CPUS=" __stringify(NR_CPUS)) : "",
Signature · 1 line
2.21.0
#4 lore raw

Le 12/07/2019 à 08:25, Michael Ellerman a écrit :

Quoted 34 lines
Quoted 32 lines · Aneesh Kumar K.V

Avoids confusion when printing Oops message like below

Faulting instruction address: 0xc00000000008bdb4

  Oops: Kernel access of bad area, sig: 11 [#1]
  LE PAGE_SIZE=64K MMU=Radix MMU=Hash SMP NR_CPUS=2048 NUMA PowerNV

This was because we never clear the MMU_FTR_HPTE_TABLE feature flag even if we run with radix translation. It was discussed that we should look at this feature flag as an indication of the capability to run hash translation and we should not clear the flag even if we run in radix translation. All the code paths check for radix_enabled() check and if found true consider we are running with radix translation. Follow the same sequence for finding the MMU translation string to be used in Oops message.

---

Changes from V1:
* Don't clear the HPTE feature flag.

  arch/powerpc/kernel/traps.c | 15 +++++++++++----
  1 file changed, 11 insertions(+), 4 deletions(-)
arch/powerpc/kernel/traps.c +0 -0
index 11caa0291254..b181d6860f28 100644
@@ -250,15 +250,22 @@ static void oops_end(unsigned long flags, struct pt_regs *regs,
  }
  NOKPROBE_SYMBOL(oops_end);
 
Quoted 11 lines

arch/powerpc/kernel/traps.c static void oops_end(unsigned long flags, struct pt_regs *regs, · line 260

Quoted 8 lines
arch/powerpc/kernel/traps.c +8 -0
+static char *get_mmu_str(void)
+{
+	if (early_radix_enabled())
+		return " MMU=Radix";
+	if (early_mmu_has_feature(MMU_FTR_HPTE_TABLE))
+		return " MMU=Hash";
+	return "";
+}

We don't change MMU once we're up, so just do this logic once and stash it into a static string, rather than rechecking on every oops.

Do we really have oops so often that we have to worry about that ?

Christophe

Quoted 19 lines

cheers

Quoted 16 lines
arch/powerpc/kernel/traps.c +2 -4
  static int __die(const char *str, struct pt_regs *regs, long err)
  {
  	printk("Oops: %s, sig: %ld [#%d]\n", str, err, ++die_counter);
 
-	printk("%s PAGE_SIZE=%luK%s%s%s%s%s%s%s %s\n",
+	printk("%s PAGE_SIZE=%luK%s%s%s%s%s%s %s\n",
  	       IS_ENABLED(CONFIG_CPU_LITTLE_ENDIAN) ? "LE" : "BE",
-	       PAGE_SIZE / 1024,
-	       early_radix_enabled() ? " MMU=Radix" : "",
-	       early_mmu_has_feature(MMU_FTR_HPTE_TABLE) ? " MMU=Hash" : "",
+	       PAGE_SIZE / 1024, get_mmu_str(),
  	       IS_ENABLED(CONFIG_PREEMPT) ? " PREEMPT" : "",
  	       IS_ENABLED(CONFIG_SMP) ? " SMP" : "",
  	       IS_ENABLED(CONFIG_SMP) ? (" NR_CPUS=" __stringify(NR_CPUS)) : "",
Signature · 1 line
2.21.0
#5 lore raw
Quoted 2 lines · Christophe Leroy

Le 12/07/2019 à 08:25, Michael Ellerman a écrit :

Quoted 1 line

"Aneesh Kumar K.V" <[email protected]> writes:

...

Quoted 20 lines
Quoted 8 lines
Quoted 8 lines
arch/powerpc/kernel/traps.c +0 -0
index 11caa0291254..b181d6860f28 100644
@@ -250,15 +250,22 @@ static void oops_end(unsigned long flags, struct pt_regs *regs,
  }
  NOKPROBE_SYMBOL(oops_end);
 
Quoted 11 lines

arch/powerpc/kernel/traps.c static void oops_end(unsigned long flags, struct pt_regs *regs, · line 260

Quoted 8 lines
arch/powerpc/kernel/traps.c +8 -0
+static char *get_mmu_str(void)
+{
+	if (early_radix_enabled())
+		return " MMU=Radix";
+	if (early_mmu_has_feature(MMU_FTR_HPTE_TABLE))
+		return " MMU=Hash";
+	return "";
+}

We don't change MMU once we're up, so just do this logic once and stash it into a static string, rather than rechecking on every oops.

Quoted 1 line

Do we really have oops so often that we have to worry about that ?

Sometimes :)

But no I don't mean it's a performance issue, it just seems simpler to compute the value once and store it. In fact for most platforms it can just be a static string at compile time, it's only 64-bit Book3S that needs to do anything at runtime.

cheers

#6 lore raw

Le 12/07/2019 à 14:22, Michael Ellerman a écrit :

Quoted 24 lines
Quoted 2 lines · Christophe Leroy

Le 12/07/2019 à 08:25, Michael Ellerman a écrit :

Quoted 1 line

"Aneesh Kumar K.V" <[email protected]> writes:

...

Quoted 20 lines
Quoted 19 lines
Quoted 6 lines
arch/powerpc/kernel/traps.c +0 -0
index 11caa0291254..b181d6860f28 100644
@@ -250,15 +250,22 @@ static void oops_end(unsigned long flags, struct pt_regs *regs,
   }

arch/powerpc/kernel/traps.c static void oops_end(unsigned long flags, struct pt_regs *regs, · line 260

Quoted 10 lines
arch/powerpc/kernel/traps.c +8 -0
   NOKPROBE_SYMBOL(oops_end);
 
+static char *get_mmu_str(void)
+{
+	if (early_radix_enabled())
+		return " MMU=Radix";
+	if (early_mmu_has_feature(MMU_FTR_HPTE_TABLE))
+		return " MMU=Hash";
+	return "";
+}

We don't change MMU once we're up, so just do this logic once and stash it into a static string, rather than rechecking on every oops.

Quoted 8 lines
Quoted 1 line

Do we really have oops so often that we have to worry about that ?

Sometimes :)

But no I don't mean it's a performance issue, it just seems simpler to compute the value once and store it. In fact for most platforms it can just be a static string at compile time, it's only 64-bit Book3S that needs to do anything at runtime.

Right, but I'm sure GCC will take care of that since the function is static and called only once.

Christophe

Quoted 3 lines

cheers

#7 lore raw
In reply to Aneesh Kumar K.V
Quoted 15 lines · Aneesh Kumar K.V

Avoids confusion when printing Oops message like below

Faulting instruction address: 0xc00000000008bdb4

 Oops: Kernel access of bad area, sig: 11 [#1]
 LE PAGE_SIZE=64K MMU=Radix MMU=Hash SMP NR_CPUS=2048 NUMA PowerNV

This was because we never clear the MMU_FTR_HPTE_TABLE feature flag even if we run with radix translation. It was discussed that we should look at this feature flag as an indication of the capability to run hash translation and we should not clear the flag even if we run in radix translation. All the code paths check for radix_enabled() check and if found true consider we are running with radix translation. Follow the same sequence for finding the MMU translation string to be used in Oops message.

Quoted 1 line

Applied to powerpc next, thanks.

https://git.kernel.org/powerpc/c/d7e02f7b7991dbe14a2acfb0e53d675cd149001c

cheers